Coffee, Myocardial Infarction, and CYP NomenclatureCoffee, Myocardial Infarction, and CYP Nomenclature

To the Editor: In their study of coffee, CYP1A2 genotype, and myocardial infarction risk, Ms Cornelis and colleagues¹ concluded that carriers of the CYP1A2*1F allele, which they define as a slow allele for caffeine metabolism, have an increased risk of nonfatal myocardial infarction compared with carriers of the CYP1A2*1A allele. The increased risk associated with coffee intake was only observed among the carriers of the *1F allele, and the effect was similar in smokers and nonsmokers.

There are 2 major problems related to this study. First, we believe that the authors' definition of the alleles is incorrect. According to the official names determined by the international CYP Allele Nomenclature Committee,² and commonly used in articles published in this research field since the CYP1A2 nomenclature was accepted in 2001, the “reference” –163 position has a cytosine and thus the wild-type (consensus) name for the CYP1A2*1A allele is “–163C.” By measuring “the –163A→C variant” (or +734A→C, as defined from the transcriptional start site² ), the authors are actually describing the consensus allele CYP1A2*1A. Thus, the CYP1A2*1F allele, as erroneously named by the authors, is instead CYP1A2*1A. This error creates exactly the opposite conclusions from those described by the authors based on this definition. Although their statement that the –163A→C nucleotide substitution creates decreased enzyme inducibility is correct, the authors generally define the –163C allele erroneously as the low-activity allele.

Second, the CYP1A2*1F allele, originally described by Sachse et al,³ has been reported to cause higher CYP1A2 enzyme induction in response to smoking³ or omeprazole treatment.⁴ However, carriers of this allele do not differ from carriers of the CYP1A2*1A allele for noninducible caffeine metabolism according to studies including those by Sachse et al⁵ and Takata et al.⁶ To our knowledge, no study has ever shown any difference in noninducible caffeine metabolism between carriers of the CYP1A2*1A vs the CYP1A2*1F allele. Therefore, the CYP1A2*1F allele cannot be defined as a “slow” or “rapid” caffeine allele as was done by Cornelis et al. If any effect of the CYP1A2*1A/1F polymorphism were seen, it would be restricted to smokers in whom an effect on inducibility might be evident, but the authors did not find this.

This confusion underscores the importance of following the international nomenclature for the human CYP alleles² and to carefully examine any true genotype-phenotype relationship described before making conclusions from epidemiologic studies. A relationship among the CYP1A2 polymorphism, caffeine metabolism, and risk of myocardial infarction remains possible but is highly improbable.